INDEX

Explanations

phrases related to positive emotional states like uplifted and uplifting

oai_token-act-pair · gpt-3.5-turbo

references to upliftment and emotional expressions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãģ®éŃĶ

-0.71

 Fiscal

-0.71

 Mistress

-0.70

 Electronic

-0.69

SHIP

-0.69

 Geneva

-0.69

PRES

-0.68

nered

-0.68

 Continued

-0.67

Han

-0.67

POSITIVE LOGITS

upl

1.13

ifted

1.13

ift

1.03

ifting

1.03

itud

1.01

atters

0.96

astically

0.95

anted

0.93

ands

0.92

ongs

0.91

Activations Density 0.015%

phrases related to positive emotional states like uplifted and uplifting

references to upliftment and emotional expressions

No Comments

No Known Activations